Mythos Sends Global Financial Elites Into Panic and Confusion: Bank of England Governor Asks, 'What Did I Do Wrong in a Past Life?'

War and historic energy shocks have already placed an unbearable burden on global financial stability. However, Anthropic's yet-to-be-released AI model, Mythos, has added another layer of unease for global policymakers.

According to media reports, during the spring meetings of the International Monetary Fund and the World Bank, military conflicts in the Middle East dominated the agenda. Governments were forced to reactivate crisis response plans, commit to coordinated cooperation as much as possible, and use fiscal and monetary tools to boost consumption. In contrast, when it came to Mythos, attendees had far more questions than answers.

European Central Bank President Christine Lagarde stated on Tuesday during a media interview that if Mythos fell into the wrong hands, the consequences would be unimaginable. She was referring specifically to Mythos's destructive potential—a concern shared by other central bank governors, finance ministers from multiple countries, regulatory officials, and heads of investment institutions. All are urgently seeking information about potential threats and protective measures.

This issue entered the public eye following a Bloomberg News report: U.S. Treasury Secretary Scott Bessent convened Wall Street leaders to warn about an upcoming AI model that could usher in a new era of robot-driven autonomous cyberattacks. On Wednesday, Bessent said, "I am confident that everyone now shares a consensus, working in the same direction to build resilience against risks."

However, outside the United States, the urgency conveyed at this meeting sparked high alertness among senior officials and bankers worldwide. They worry that this technology could breach traditional cyber defenses, exposing the financial system to unpredictable threats.

Earlier this week, Bank of England Governor Andrew Bailey, speaking at an event at Columbia University in New York, remarked, "Logically, the situation in the Gulf region should be our latest challenge. But then you wake up one day and realize Anthropic may have found a way to unlock the entire world of cyber risks—and you wonder: What did I do wrong in a past life?"

Bailey and other top experts in financial stability face a dilemma: their understanding of related threats remains limited, making it impossible to assess how far Mythos has advanced beyond the cybersecurity risks they have been warning about for years. Many questions remain unresolved:

Will Mythos trigger large-scale thefts of bank accounts, a total paralysis of international payment systems, or even a confidence crisis that shakes the very foundation of the financial system?

Is this similar to the risk posed by quantum computing—a known threat regulators are already studying deeply—or does it require entirely new response strategies?

It remains unclear to what extent San Francisco-based Anthropic is sharing details about Mythos and granting access to non-U.S. officials and bankers. A U.S. executive leading a European bank stated that they have received assurances of briefings. Restricted access could leave parts of the global financial system unable to follow the IMF's recommendations to stay ahead of evolving cyber threats.

Canadian Finance Minister François-Philippe Champagne emphasized that Mythos demands full attention and expressed his desire to consult with counterparts from other nations on the issue. "We share common interests and must ensure the resilience of the financial system," he said.

According to media reports, a technical team from the U.S. Treasury Department is seeking access to Anthropic's Mythos AI model to begin identifying potential vulnerabilities.

The Federal Reserve Bank of New York serves as the core hub for global financial transactions and the regulation of stability risks. On Thursday, Fed President Williams noted that recent technological advances have fully demonstrated the powerful capabilities of these AI tools—whether in identifying vulnerabilities or exploiting them—with development speeds far exceeding many expectations.

Citing sources familiar with the matter, media outlets reported that at Wednesday's G7 finance ministers' meeting, participants discussed the necessity of establishing an international institutional framework to govern AI. The consensus was clear: no single country can handle the risks posed by Mythos alone. One source added that given the exponential growth of risks and potentially severe consequences, attendees generally supported deeper exploration of the issue, though specific pathways remain unclear.

To make matters worse, this doomsday AI scenario coincides with a sharp rise in geopolitical risks. At such a point, technology's role would shift from being an engine boosting financial sector efficiency to becoming a weapon capable of destroying its own technological infrastructure.

Industry insiders point out that the pace of development in frontier AI capabilities has already outstripped the speed of designing governance frameworks for them. Before these models become deeply embedded in critical systems like finance, building safeguards carries genuine urgency.

Media analysis suggests that this week's pursuit of answers also highlights the growing divide between "haves" and "have-nots" in the AI field, as well as the increasing fragmentation of the global financial stability ecosystem. European concerns over excessive reliance on foreign tech companies continue to grow.

Privately, officials also express deep concern over the increasingly prevalent "go-it-alone" tendency in global financial dialogues, fearing it will hinder effective sharing of risk information worldwide. Several European officials stated they would use this week's meetings as an opportunity to urge their American counterparts to share relevant information as extensively as possible.

Swedish Finance Minister Elisabeth Svantesson announced that later Thursday there would be a warning session with central bank governors and finance ministers, with AI as one of the key topics. "My message is clear: AI is wonderful, but combining it with unpredictable leadership is dangerous. The war between the United States and Israel, along with uncertainty and cyber threats, are the dominant issues today," she said.